Google: Gemma 4 31B (free)
Google: Gemma 4 31B (free) is the free-tier variant of google's Gemma 4 31B model, available at $0 per million tokens for both input and output. It supports multimodal inputs including text, images, and video, with a 262,144-token context window. This variant has not been included in our benchmark testing suite, so we do not have per-dimension scores or an overall ranking specific to the free tier. The paid Gemma 4 31B has been tested and scored an average of 4.416 across our 12-benchmark suite — placing it in the upper tier of tested models, above all other non-free google models except Gemini 3 Flash Preview (avg 4.5) and Gemini 3.1 Flash Lite Preview (avg 4.416). The free variant provides zero-cost access to the same underlying model, typically with rate limits applied.
Performance
We have not benchmarked Google: Gemma 4 31B (free) directly in our 12-test suite, so per-benchmark scores and rankings are not available for this variant. The paid Gemma 4 31B has been tested and scored an average of 4.416 across our benchmarks — one of the stronger scores in the tested set, tied with Gemini 3.1 Flash Lite Preview and above most other models in the google family. If you need benchmark data to evaluate performance characteristics, the Gemma 4 31B paid profile contains full benchmark breakdowns. The free variant's performance is expected to mirror the paid version at the model level, with the primary differences being rate limits and availability.
Pricing
This model is free — $0 per million input and output tokens. Free-tier access typically includes rate limits and lower queuing priority compared to paid access. For teams evaluating the model before committing to production use, the free tier provides a zero-cost evaluation path. The paid Gemma 4 31B variant costs $0.38/M output — at 10 million output tokens per month, that's $3.80; at 100 million tokens, $38.00. For high-throughput or production workloads, the paid variant is the appropriate path. The Gemma 4 26B A4B (free) is the other free google model in the tested set, offering a similar zero-cost entry point to a comparable architecture.
Google: Gemma 4 31B (free)
External Benchmarks
Pricing
Input
$0.00/MTok
Output
$0.00/MTok
modelpicker.net
Real-World Costs
Pricing vs Performance
Output cost per million tokens (log scale) vs average score across our 12 internal benchmarks
Try It
from openai import OpenAI
client = OpenAI(
base_url="https://openrouter.ai/api/v1",
api_key="YOUR_OPENROUTER_KEY",
)
response = client.chat.completions.create(
model="google/gemma-4-31b-it:free",
messages=[
{"role": "user", "content": "Hello, Google: Gemma 4 31B (free)!"}
],
)
print(response.choices[0].message.content)Recommendation
Google: Gemma 4 31B (free) is an excellent starting point for developers who want to evaluate a high-performing multimodal model at zero cost. With the paid Gemma 4 31B scoring 4.416 on average in our testing — one of the stronger results in the tested set — the free variant gives access to a capable model for prototyping, testing, and low-volume use. The multimodal support (text, image, video) with a 262K context window makes it valuable for exploring multimodal workflows before committing to paid access. For production deployments requiring consistent throughput and rate limit guarantees, use the paid Gemma 4 31B at $0.38/M output. Teams looking for the free google model with the highest benchmark-validated performance should note that Gemma 4 31B scores higher on average (4.416) than the also-free Gemma 4 26B A4B variant's paid counterpart (4.25).
How We Test
We test every model against our 12-benchmark suite covering tool calling, agentic planning, creative problem solving, safety calibration, and more. Each test is scored 1–5 by an LLM judge. Read our full methodology.